• by Phil Leggiere , May 9, 2007

"Build it and they will come" is the classic fallacy of short-sighted marketers. The online world may have found its equivalent in the current widespread and spurious notion "drive traffic to your Web site and they will buy." Actually, as Brent Hieggelke, vice president of strategic marketing at Omniture, explains below, targeting prospects off-site through advanced behavioral methods in order to drive them on-site is only half the goal. The other half is using behavioral analysis to model what consumers do once they reach your site.

Behavioral Insider: How do the on-site behavioral targeting tools developed by Touch Clarity, now a part of Omniture, work?

Brent Hieggelke: We find that visitors tend to revisit a site as many as five times before reaching a buying decision. So what our on-site targeting does is start building a profile of each visitor as soon as they reach the site, and then subsequently recognize them using an anonymous first-party cookie. Fully automated, the software learns in real time which content will best achieve a given goal for each visitor and triggers the serving of that content.

BI: What types of data do you focus on?

Hieggelke: Personally identifiable information is not where you get your insights. Knowing your name doesn't necessarily give you anything predictive. It's really anonymous behaviors which, taken in relation, can form patterns which are predictive. Profiles are used by the software to learn which content is the most relevant to present to each individual. Each time a visitor returns, any new data is combined with retrieved historical information to ensure that the very latest and most complete view of the visitor is used as the basis for content-targeting decisions.

BI: What are the challenges you've found in educating your customers about the importance of what you call on-site targeting?

Hieggelke: We challenge our clients with this question. We say, 'You sell 200 or 300 or 500 products. Yet the homepage your customers see when they come to your site shows the same four products to every single visitor interchangeably regardless of their interests. Why?'

BI: What do they say?

Hieggelke: The standard reply is because that's all that's possible. But we ask them then, what if you could deploy only those products on your home page that you know, based on data about those incoming customers, will most likely suit their interests best. They get that concept immediately.

BI: What are the biggest gaps in publisher's conceptions of the role behavioral data can play?

Hieggelke: BT has been not so much misunderstood, but only partially understood to mean studying intensely what people do off your site. Advertisers use thousands of variables to get a predictive model of offsite behavior to help them figure out exactly which ad content to serve. But they focus all their attention on drawing customers to their site. But then they essentially stop when customers get to the site.

But say you're a consumer electronics retailer. You've employed your targeting well to address people who've shown by their behavior surfing and browsing theWweb that they're likely to be good prospects. You've targeted them with ads. That's a good start. And then that consumer comes to your site and spends several days looking at big-screen TVs. Why would you want them to see car stereo information the next time they come to your site?

BI: How does what you're doing integrate with sales automation and customer relationship management tools?

Hieggelke: Web sites have managed to disintermediate much of the sales process from salespeople. By the time consumers in many verticals reach 'point of sales.' whether it's online or ultimately in a store, they've done 95% of their research already. But what's been missing during that 95% of the time is the ability to produce more of an actively shaping sales presence within the We bsite.

BI: Which verticals have been the most eager adopters of this approach?

Hieggelke: We've gotten very strong adoption in financial services. That's an area, for instance, where there's a great value placed on offering a wide variety of portfolio products for cross-selling and up-selling customers. Retail is also an area where onsite analytics can help leverage stores with a deep catalog of products. We see this approach as useful for direct-response marketers, certainly -- but perhaps even more so for brands which have invested enormous amounts of money to make their sites a true destination.

We produce reports documenting the uplift on targeted onsite customers versus control groups with no targeting, and the differences are dramatic. We're regularly seeing improvements of well over 100% in post-click conversions.

BI: Do you see the type of targeting you're advancing as being readily adaptable to other channels?

Hieggelke: The key thing is to automate the process so that clients can concentrate their energies on marketing. We don't serve content. What we do is offer the tools for making decisions about what kind of content to present to particular consumers. The type of channel doesn't really matter. Right now it's the Web site or the Web kiosk, but it could be mobile messaging, and it could be video-on-demand cable.